summaryrefslogtreecommitdiffstatshomepage
path: root/src/mame/drivers/pgm2.cpp
diff options
context:
space:
mode:
author MetalliC <0vetal0@gmail.com>2017-12-30 09:11:22 +0200
committer MetalliC <0vetal0@gmail.com>2017-12-30 09:11:54 +0200
commit94d584d214bcd8f5b6e9e7b7f9c5779bc67c7892 (patch)
treef690bd2d4b62353215128463fa6d1d405f1fb39d /src/mame/drivers/pgm2.cpp
parentc5203b2540a4cef67f88920335dfbcad7dfa2247 (diff)
pgm2.cpp: add note (nw)
Diffstat (limited to 'src/mame/drivers/pgm2.cpp')
-rw-r--r--src/mame/drivers/pgm2.cpp5
1 files changed, 5 insertions, 0 deletions
diff --git a/src/mame/drivers/pgm2.cpp b/src/mame/drivers/pgm2.cpp
index f3f93471512..bd29ee1e3e3 100644
--- a/src/mame/drivers/pgm2.cpp
+++ b/src/mame/drivers/pgm2.cpp
@@ -364,6 +364,11 @@ WRITE16_MEMBER(pgm2_state::unk30120014_w)
It is not clear if real address/data xor values derived from written "key",
or FPGA just waiting to be be written magic value at specific address in ROM area, and if this happen enable descrambling using hardcoded values.
+
+ Current implementation assume "expected sum" read at step 6) is regular data from ROM (descrambled),
+ our keys calculated assuming this, so further ROM sum check passes OK.
+ But, there is chances in hardware its not like that, but these bytes comes from FPGA (some indication of this is the fact 5) and 6) is actually single routine).
+ In this case we have 2 unknowns (key and summ) and can not guess both. Will be needed serial comm logs from real hardware.
*/
READ_LINE_MEMBER(pgm2_state::module_data_r)